UCAR > Communications > Quarterly > Fall 1996 Search

Unidata and the universities:
Have we been here before?

by Otis Brown
Chair, Unidata Policy Committee, and Dean, Rosenstiel School of Marine and Atmospheric Science, University of Miami

The Unidata Program is now planning for the future: What should Unidata be like in 2003? What will the universities need from the program by then? In pondering these questions, I reread two articles on Unidata published in this newsletter earlier this decade. Both were written by Robert Fox, who is deputy director of the University of Wisconsin-Madison's Space Science and Engineering Center and was then chairman of the Unidata Policy Committee. And both were instructive.

In his 1990 article, Fox recounted the early history of the program and outlined its accomplishments in the ten years since its inception. Reading the article, I was struck by how close are the problems we in the universities currently face to those that spawned the birth of Unidata.

The second article was written in 1992, when Unidata's proposal for continued funding was being reviewed by NSF. In addition to a summary of the program's recent progress, the article focused on the challenges then facing the program. Here it is interesting to compare the perceived challenges with the reality of the past five years.

As Fox noted in 1990, the Unidata Program was created to solve two problems: "Access to and acquisition of atmospheric science data, and access to interactive tools for conducting education and research with those data." Today, rather than having a dearth of data or tools, the universities are on the brink of being overwhelmed by both. New observing systems, including new satellites such as the Geostationary Operational Environmental Satellite (GOES-NEXT), the Earth Observing System, and systems implemented in the course of the NWS modernization, are beginning to create a "fire hose" of data. This fire hose was perceived as a challenge in 1992, but only recently have we come to realize that providing "access to data" really translates into devising methods for sorting, storing, and displaying very large volumes of data. This means creating new decoders at a minimum, and probably new display packages will also need to be created, distributed, and maintained. Who is going to undertake these tasks and how are they going to be funded? Which data take priority and who is going to decide?

On the technology front, Unidata and the universities must continually adapt to changes in both computer networking and in computing platforms. For most of Unidata's history, the university community relied on NSFnet to provide cheap and reliable connectivity. In this environment, the Unidata Program developed its Internet Data Distribution (IDD) system, which uses computer networks for data distribution, replacing the more expensive satellite-broadcast-based distribution method. The challenge in 1992 was making sure the newly conceived IDD would really work.

Last year, however, NSF unplugged NSFnet, beginning the five-year transition of the Internet from a federally funded activity to a private-industry endeavor. This transition, coupled with the discovery of the Internet by the general public, has caused such dislocations that there are now doomsayers predicting the death of the Internet by the end of this year. None of this was foreseen in 1992.

While I do not believe the Internet is in any danger of disappearing, I do believe that university access and use of the Internet may soon change fundamentally. Communications giants such as AT&T, MCI, and others are building private, high-speed backbones and offering their services under market conditions. Where do the universities fit in this scheme? How many of us are capable of paying, or planning to pay, for the quality of service and bandwidth we will soon need? How does this affect Unidata's IDD?

At the same time, the computer industry itself is undergoing paroxysms: computer memory has become relatively inexpensive, platform performance has improved manyfold, and operating systems are evolving rapidly. University purchases of equipment currently are strongly influenced by Unidata's support of only the UNIX and OS/2 operating systems. How long will these be appropriate choices? Should Unidata (hence universities) move to WindowsNT, or will the promise of networks and new computer languages, such as Java, result in our becoming operating-system independent? And how do we make the transition in either case?

Essentially, then, we've come full circle: we're facing a new generation of the same issues we faced in the early 1980s. And, as usual, we lack a crystal ball. How many of us back then foresaw the effects of NSFnet? Even in 1992, only spiders wove webs and java meant a cup of coffee. What new technology will appear in the next five years to force on us yet another paradigm shift?

As we face this brave new world yet again, there is one important change: this time the air of crisis is missing. I credit the existence of Unidata for this. The Unidata paradigm of broad university-based governance and decision making with centralized management and support of limited tools has proven to be highly successful in meeting the challenges of adapting to an evolving technological environment.

But I fear complacency. The challenges now facing us are at least an order of magnitude more complicated than previously, and meeting them will require a concerted effort on the part of not only Unidata, but each and every university. The Unidata Program Center will be submitting a proposal to NSF next spring and is now beginning to develop a strategy for accomplishing its mission through 2003. Unidata staff members may be contacting you to solicit information for this proposal. I strongly encourage you to share your views on the past performance and future goals of the program. Comments may be directed to me (Rosenstiel School, University of Miami, 4600 Rickenbacker Causeway, Miami, FL 33149; 305-361-4000 or obrown@rsmas/.miami.edu). You may also contact David Fulker, director of the Unidata Program Center (303-497-8650 or fulker@ucar.edu), or Sally Bates, Unidata information manager (303-497-8637 or sally@unidata.ucar.edu).

In this issue... Other issues of UCAR Quarterly

Edited by Carol Rasmussen, carolr@ucar.edu
Prepared for the Web by Jacque Marshall
Last revised: Tue Apr 4 10:18:15 MDT 2000